基于半分析模型的高光谱遥感水深探测方法  

Bathymetry Using Hyperspectral Imagery Based on Semi-analytical Model

作　　者：张靓[1] 滕惠忠[1] 孟婵媛[1] 李军[1] 辛宪会[1] 叶秋果[1] 李海滨[1] 

机构地区：[1]海军海洋测绘研究所,天津300061

出　　处：《海洋测绘》2011年第4期17-21,共5页Hydrographic Surveying and Charting

基　　金：国家863计划项目(2009AA121403)

摘　　要：海岸带浅水和底质的成分复杂且分布不均,使遥感水深探测面临很大挑战。尝试通过利用Hydrolight仿真建立的遥感反射率半分析模型来解决这个问题。采用英属维尔京群岛彼得岛的高光谱影像,应用非线性优化的方法提取了水下地形。从反演水深和海图标注水深拟合的趋势线斜率为0.9392,决定系数为0.8258。试验结果证明,使用定标好的高光谱影像进行海岸带地区快速水深探测是可行的。The complexity and heterogeneity of shallow coastal waters and benthic habitats provide a challenge for bathymetry using remote sensing imagery. We attempt to overcome this problem by utilizing a semianalytical model of remote sensing reflectance, generated through Hydrolight silmulation. This study applied a nonlinear optimisation methodology to retrieve underwater terrain using hyperspectral imagery of Peter Island in the British Virgin Islands. The slope of the regression line derived from the retrieved bathymetry and the depths of the chart was 0. 9392, while the coefficient of determination is 0. 8258. These results support the use of calibrated hyperspectral imagery for the rapid determination of bathymetry of coastal environments.

关 键 词：高光谱遥感 水深探测 半分析模型 非线性优化 

分 类 号：P237[天文地球—摄影测量与遥感]